IREE runtime in Godot through GDExtension, a mission to run a machine learning model (e.g. Tensorflow lite) natively in Godot.
Authored by Richie Kho and its contributors.
| Platform | HAL Backend used |
|---|---|
| Apple products (macOS, IOS) | metal |
| Desktops (Windows, Linux, *BSD, Android) | vulkan |
| The rest | vmvx |
The documentation is hosted using Github page.
The sample project is in sample directory.
We'll use Git and CMake to build this project.
git clone https://github.com/iree-gd/iree.gd.git # clone this repo
cd iree.gd
git submodule init thirdparty # initialize all the thirdparty
git submodule deinit thirdparty/iree/third_party/llvm-project # Deinitialize LLVM, as we are not compiling the compiler.
git submodule update --recursive # Pull submodule content, this will take a while.
mkdir build
cd build
cmake ..
cmake --build .If you would like to compile LLVM from the source, you'll need to set the IREE_BUILD_BUNDLED_LLVM CMake option to ON when generating build files with CMake and also initialize the llvm-project submodule under the thirdparty/iree/third_party/llvm-project.
After compilation, the library will be in build/lib directory.
If the COPY_LIBS_TO_SAMPLE CMake option is ON, the library will also be installed into the sample.
